Overview
Designed an automated Weekly Product Review Pulse that transforms public App Store and Google Play reviews into a concise one-page insight report for fintech products. The system clusters review feedback, surfaces recurring themes, pulls representative quotes, and highlights actionable ideas to help teams quickly understand customer sentiment. Reports are delivered through Google Workspace using MCP servers for Google Docs and Gmail, keeping the workflow structured, repeatable, and easy to share.
AI Tools
Stitch by Google
Cursor
Objective
Give product, support, and leadership teams a repeatable, weekly snapshot of what customers are saying in store reviews: themes, representative quotes, and actionable ideas without manual copy-paste or one-off spreadsheets.
-
Build an AI Agent to ingest, analyze, and summarize customer reviews into actionable weekly insights.
-
Develop MCP Servers for Google Docs and Gmail to automatically generate reports, maintain historical records, and distribute stakeholder updates.
-
Automate customer feedback analysis by clustering reviews, identifying recurring themes, and surfacing actionable product opportunities without manual effort.
What the system does
-
Ingest public reviews from the last 8–12 weeks (configurable window) from Google Play for the GROWW product.
-
Use an LLM to name themes, pull verbatim quotes, and propose action ideas—with validation so that quotes must appear in real review text.
-
Render a concise one-page narrative: top themes, quotes, action ideas, and a short “whom this helps” section.
-
Deliver outputs only through Google Workspace MCP servers:
-
Google Docs MCP — append each week’s report as a new dated section to a single running document per product (e.g., Weekly Review Pulse — Groww). The Doc is the system of record and preserves history.
-
Gmail MCP — send a short stakeholder email that includes a deep link to the new section in that Doc (heading link), not a duplicate full report in the email alone.
-
What this solves
Automates the process of turning customer reviews into actionable insights, helping product teams prioritize roadmap decisions, support teams identify recurring issues, and leadership gain a clear weekly snapshot. Reports are automatically generated, archived in Google Docs, and shared via Gmail with a direct link to the latest report.
Building with Cursor
I used Cursor to build the workflow end to end, from structuring the review pipeline to refining the automation logic.
To keep the system modular, I created a separate MCP server in GitHub to handle the Google Docs and Gmail draft automation independently. This allowed me to generate the report in Docs, prepare the email draft in Gmail, and keep the workflow clean, reusable, and easy to maintain.
Final UI — Refined Experience
I designed the user interface in Stitch, focusing on a clean and intuitive chatbot experience. I refined the layouts, interactions, and response patterns to ensure clarity and ease of use. Once the interface was integrated and tested, I deployed the application on Vercel, making the end-to-end prototype accessible and ready for demonstration.
→ View app
